HHS AI Chief Calls for Collaboration in Strategy Execution

The Department of Health and Human Services’ Artificial Intelligence Strategy is driving its four focus areas forward with a central theme of collaboration, the agency’s Chief AI Officer Oki Mek said during NVIDIA’s virtual event last week.

Released in January, the strategy aims to enhance the health and wellbeing of all Americans. The strategy has four main focuses, which include forming an AI Council, AI Community of Practice, governance and compliance, as well as support for research and development in AI.

Across those four focus areas, Mek said collaboration and idea-sharing are the best approaches his agency can take in accomplishing its mission with AI. He emphasized this point especially in the first focus of the AI Strategy — the formation of the AI Council.

“I believe AI is a team sport,” Mek said. “We want each operating division and staff division across the department to have a voice with how the department is carrying out the AI strategy. We believe in collaboration. We stand to learn a lot from one another. Since AI is not confirmed to just technology, the more I learn, the more I realize how much I don’t know much in this domain.”

Collaboration also means focusing on inclusion, in that different people have different perspectives on privacy, data, security and human-centered AI, Mek said, highlighting the AI Community of Practice’s cooperative nature.

“We want to celebrate AI initiatives, share lessons learned, promote best practices, establish an AI playbook that is targeted at the health sector and focus on two biggest obstacles in AI: data acquisition or collection and data preparation,” Mek said.

The third focus, establishing governance and compliance around AI, will ensure that HHS does not violate any policies or regulations around privacy, ethics, legality, code bias, IT and IT security, as well as acquisition. Then the research and development focus will create a testbed to lower risk for HHS personnel to explore AI.

These last two are also collaborative in nature, contributing to HHS’s “spoke and wheels” centralized model for AI development, Mek said.

“We want to change the mindset from, ‘We don’t want to share data’ mantra to ‘we must share data,’” Mek said. “There has to be [a] hard justification why data is not being shared. Without data, there’s no AI.”

Mek stressed that as HHS executes the strategy, the agency must continue to push for transparency, trustworthiness, bias prevention and data provenance across all its AI initiatives, since the strategy and collaboration behind it must ultimately serve the public interest.

“What we do with AI must benefit society,” Mek said.